著者
北村 泰彦 寺西 憲一 辰巳 昭治 Yasuhiko Kitamura Ken-ichi Teranishi Shoji Tatsumi
雑誌
人工知能学会誌 = Journal of Japanese Society for Artificial Intelligence (ISSN:09128085)
巻号頁・発行日
vol.11, no.3, pp.470-477, 1996-05-01
参考文献数
9
被引用文献数
7

Real-Time A^*(RTA^*) is an on-line search scheme in which look-ahead searches and moves are interleaved. It does not guarantee to find an optimal solution but it finds a semi-optimal solution with less search effort. To improve the quality of solution, Knight has proposed Multi-Agent Real Time A^*(MARTA^*), in which multiple agents concurrently and autonomously execute RTA^* for a given problem. In this scheme, agents do not coordinate their moves, but each agent just randomly chooses its move when the candidates look equally good. In this paper, we have interest in how agents should coordinate their moves to find a better solution faster. We propose two organizational approaches based on scattering and gathering. Each agent measures the distances from other agents and chooses its move in the direction of departing or approaching. These two approaches strengthen the discovery effect to find more undiscovered solutions and the learning effect to find more good solutions of MARTA^* respectively. We evaluate them through simulation experiments on maze and 15-puzzle problems and analyze why they work well or not from a point of heuristic depression, which is a local hollow of heuristic state evaluation values on paths to goal states. Once an agent falls into a depression, it cannot escape without filling in the depression, namely updating the state evaluation values. In a maze problem, in which deep depressions are scattered in its state space, scattered agents show better performance than gathered agents because less agents fall into depressions. On the other hand, in a 15-puzzle problem, in which shallow depressions are ubiquitous, gathered agents show better performance because they are better at cooperating to fill in the depressions. As a result, it is shown that we can get better search performance in MARTA^* by using an appropriate organization of agents depending on the characteristic of heuristic depressions of the given problem.
著者
植野 研
出版者
一般社団法人 人工知能学会
雑誌
人工知能学会誌 = Journal of Japanese Society for Artificial Intelligence (ISSN:09128085)
巻号頁・発行日
vol.21, no.1, 2006-01-01

演奏家や体操選手などの技芸実践者は, 長い年月をかけて高度な運動スキルを獲得し, 通常は成し遂げることのできない技を実演することができる.しかしながら, 多大な労力と時間をかけて磨き上げた技芸スキルは, 一般に, 個人差や筋骨格の冗長性・流派などの違いから, 言葉で説明することが困難であり, これを正しく理解することが困難であることが知られている.本研究では, このような身体知の構築問題に対し, 身体制御のタイミングに着目し, 高度なスキルをピークタイングシナジー(PTS)としてモデル化する方法を提案する.PTSとは, 身体各部のタイミングの協調制御に関する動作一貫性制約で, 運動タスクにおいて遵守すべき身体知のことである.次に, このPTSを, 身体計測データから自動構築するアルゴリズムPRESTO (Phasic Relation Extraction for Skill Training and Optimization)を提案する.PRESTOは, 多角形近似を用いて波形の変化ピークとその時間ずれを特徴点とし, 時系列マイニングによりタイミングの法則性を導き出す.本方法により, 多変量時系列波形から身体各部の連鎖動作を点と線で捉えるPTSのモデル化が可能となる.本論文では, チェロ演奏における運弓動作をとりあげ, 計測データからPTSを構築する検証実験を行った.この検証実験により, 熟練者に共通する共通スキル, 特定の熟練者に特化した個性スキルをモデル化でき, 解釈可能な身体知を構築することを確認した.技芸スキル教育におけるスキル解析の一つの道具立てを提供することができ, 技芸スキル理解における足がかりをつけることができた点で, 本研究は大きな意義をもっていると考えられる.
著者
大島 千佳
出版者
一般社団法人 人工知能学会
雑誌
人工知能学会誌 = Journal of Japanese Society for Artificial Intelligence (ISSN:09128085)
巻号頁・発行日
vol.20, no.1, 2005-01-01

本研究は,クラシック音楽に見られるような「再現演奏」を行う奏者が,作品に対する内的感動をより良く表現できるようにすることを目的とする.まずピアノレッスンのケーススタディを行い,生徒の演奏データ(MIDIデータ)を生徒のブラインドによる主観評価と照らし合わせた分析を試みた.その結果,主観評価では明示されていない生徒の音楽的理解度や技術習得上の問題点を推定することができた.次に,MIDIシーケンスデータの「2段階式作成システム」を構築し,音楽表情に関わらない「音高の再現」をシステムが直接に支援することで,音楽表情の表出を間接的に支援できることを示した.さらに,このシステムを応用して誰でもすぐにピアノ連弾ができる「Family Ensemble」を提案した.また,複数の奏者による演奏構築支援を目的に,ピアノ連弾の練習中の対話で使用された「楽器奏」の機能を基盤化の理論を用いて分析した.本論文では,奏者に対しての何らかの支援を目的としているが,音楽表情を高めることを阻害するような行き過ぎた支援を行わないように特に配慮している.そのためには「音楽」を知るとともに,多くの分野の学問を結集させることが必要であった.このように本論文は,各分野の知識を結集した知識科学的なアプローチによって,演奏構築における音楽表情の形成過程の解明と支援に取り組んだ最初の試みとして位置づけられよう.

1 0 0 0 OA 会議報告

雑誌
人工知能学会誌 = Journal of Japanese Society for Artificial Intelligence (ISSN:09128085)
巻号頁・発行日
vol.28, no.4, pp.676-680, 2013-07-01
著者
小方 孝 堀 浩一 大須賀 節雄 Takashi Ogata Koichi Hori Setsuo Ohsuga
雑誌
人工知能学会誌 = Journal of Japanese Society for Artificial Intelligence (ISSN:09128085)
巻号頁・発行日
vol.11, no.1, pp.148-159, 1996-01-01

In this paper, we describe a basic framework of the narrative generation system for supporting human creative tasks. The narrative generation process by computer is divided into the conceptual representation level and the surface language generation level, and we deal with only the former level here. The conceptual representation is divided into three aspects; story, plot, and construction. While the story is an events sequence that was arranged according to a temporal order, the plot is an events sequence that was reorganized by an order which each event is introduced into a narrative. These three levels in a narrative are constructed as tree structures. Terminal nodes in the tree structures are events and all nodes other than them are relations that connect subordinate nodes. Narrative generation is performed by expanding or transforming a tree structure. In the story and construction generation, the system enlarges each tree by expanding events or partial trees using appropriate relations, and in the plot generation, a story tree is transformed into a plot tree through the connection relations among nodes in it are rearranged. We call narrative techniques the procedures to expand a tree through applying relations to nodes or to transform a tree using actors" viewpoints or plot patterns. On the other hand, we call narrative strategies the rules to decide a current executable narrative technique and the node to which it is applyed according to narrative parameters that define the features of a narrative to be generated through narrative generation process. The system generates a narrative by executing appropriate narrative techniques under the control of narrative strategies based on a set of events and narrative parameters were given by user. This narrative generation mechanism has some remarkable characteristics. First, the system can flexibly generate a variety of narratives from one input. Next, the system has an ability that integrates a variety of theories or knowledge representations and that extends the system itself. These advantages are relate to clear separation among narrative techniques, narrative strategies, and knowledge base. Lastly, by above reason, the system has potentiality that can use for various purposes. We can change or add each modules in it to apply to specific areas.
著者
渡辺 澄夫 Sumio Watanabe
出版者
人工知能学会
雑誌
人工知能学会誌 = Journal of Japanese Society for Artificial Intelligence (ISSN:09128085)
巻号頁・発行日
vol.16, no.2, pp.308-315, 2001-03-01
参考文献数
25
被引用文献数
20

The parameter space of a hierarchical learning machine is not a Riemannian manifold since the rank of the Fisher information metric depends on the parameter.In the previous paper, we proved that the stochastic complexity is asymptotically equal to λ log n-(m-1)log log n, where λ is a rational number, m is a natural number, and n is the number of empirical samples.Also we proved that both λ and m are calculated by resolution of singularties.However, both λ and m depend on the parameter representation and the size of the true distribution.In this paper, we study Jeffreys' prior distribution which is coordinate free, and prove that 2λ is equal to the dimension of the parameter set and m=1 independently of the parameter representation and singularities.This fact indicated that Jeffreys' prior is useful in model selection and knowledge discovery, in spite that it makes the prediction error to be larger than positive distributions.
著者
ビンステッド キム 滝澤 修 Kim Binsted Osamu Takizawa
雑誌
人工知能学会誌 = Journal of Japanese Society for Artificial Intelligence (ISSN:09128085)
巻号頁・発行日
vol.13, no.6, pp.920-927, 1998-11-01

We have implemented a simple model of puns in a program (BOKE) which generates puns in Japanese, using linguistic information from a general-purpose lexicon. Our rough evaluation indicates that the puns generated by the program are of comparable quality to those generated by humans. BOKE differs from an earlier English-language system (JAPE) only in the lexicon and the templates used to generate the surface text-the punning mechanisms are the same. This suggests that our model of puns is language independent.